There’s a difference between the real-ish-time weather data continuously fed in to output predictions, and the decades of weather data used to build the model. The continuous feed of data is more than likely part of what Google alleges is saving significant energy.
Its the training on decades of information, and occasional updates to those trained models that take a significant amount of resources, but hopefully for relatively short bursts.
I feel this personally, I live in the hills outside of a valley metro. All weather data is forecasted off of valley sensors, but shit gets weird when you suddenly climb 2000+ ft.
The best weather services in my area are those that can factor in peoples household meters into their forecasting, but those services still aren’t perfect.
It’s not just about cutting costs, but also improving accuracy. Physical simulations factor in a dozen or so weather conditions to predict outcomes. Machine learning can track thousands of conditions, drawing connections not realized in physical models, leading to much more accurate statistical models.
Why wouldn’t you eat a cat? It’s less social than a cow or a pig.
Personally, I’m too lazy to hunt, so I just get my meat from Elwood’s (https://www.elwooddogmeat.com/). Hoping they add cat to their lineup one day.
What they leave off is how much goes into training the model, but I imagine once they settle on a trained model it can carry on pretty efficiently for a long time, especially if they’re baking in things like atmospheric CO2 levels to help keep forecasts in line with global warming.
+ meat
Yeah, I said it.
Why? Every step you move up the food chain requires roughly 10x as much inputs as outputs. To get a pound of protein from a cow you have to feed it 10lbs of plant protein. Almost all cattle feed comes from farms, just like your veggies. Anywhere we grow soybeans and hay for cattle could easily be converted to growing fruits and vegetables for human consumption. There’s a small loss of efficiency by growing human-quality food instead of cattle food in these spaces, but its nothing in comparison to the loss of resources from trying to raise cattle.
Almost none of the meat we eat is truly free-range - it all gets fed farmed produce that comes from farms that could grow food for humans in a fraction of the space.
Which part are you dubious about?
Not enough land for meat? There is technically… but it requires factory farms and STILL we need to chop down huge swaths of the amazon to keep the machine churning.
Not enough land for plant based diets? Is takes only about 1/10th the amount of land to grow plants for human consumption than to grow plants for feeding our animals that we then consume. Sure not every alfalfa farm on the planet can switch to cucumbers out the gate, but well over 10% can…
Businesses make decisions based on money. People make decisions based on vibes.
Phew, glad my psychiatrist prescribed me Ritalin.
Which would be a pull, not a push.
Some people really don’t grasp the concept of the trolley problem.
A train is headed towards 5 people, you can pull a lever and switch tracks to kill 1 person. What do you do?
Naïve/entitled people fail to grasp the concept of compromise.
Mmm, I see, I see, so you’re voting for the only other person who could possibly win this election, Kamala Harris, right?
Ahh yes, of course, of course, so Trump to the rescue then right?
Oh okay, then you’re voting for Harris, right?
https://www.reddit.com/r/VanLife/comments/1avrsbd/rooftop_solar_tracker/
Looks like someone out there had the same idea.
They’re currently selling a packaged system for ~$5k, which is a bit pricey, but maybe worth if you’re full time on the van life. Looks like you could probably fit 2 of these units on a typical roof, obstructions permitting. Alternatively you could do one tracker and one flat unit.
It looks like ground based 2-axis tracking systems can be found for ~$500-1k, you’d just need to bring your own expertise to mount it to the roof and program the controls yourself.
Semi-unrelated: others mentioned painting your van white (which you already have) but I’ll raise you one better - consider painting it with an IR-Emissive cooling paint, e.g., https://i2cool.com/
There’s unfortunately few products on the market at this time, but keep your eyes peeled.
WaPo - “Democracy dies in darkness”
Bezos - “I can deliver darkness to everyone in America in 2 days or less”
They were talking about Nova, not Lawnchair. Nova hasn’t officially been abandoned, but they were purchased by a big data broker a few years ago, and just a few months ago Branch (Nova’s owner) laid off almost all of the Nova Launcher development team.
Nova is not dead, but the writing is on the wall.
It would be amazing if it could have a significant impact on spatial and temporal accuracy of things like rain. I feel like for me the existing weather report is good enough for “it will probably rain tomorrow” but it’s really hit-or-miss when you get to hourly resolution. A good model may be able to go so far as to say “it will probably rain between 3-4pm on the east side of town tomorrow, and 2-3pm on the west side”
That’s the dream at least. With enough data and a sophisticated enough model it feels like it could be possible.